How to perform mirror therapy after stroke? Evidence from a meta-analysis

Abstract

Background:

A recently updated Cochrane review for mirror therapy (MT) showed a high level of evidence in the treatment of hemiparesis after stroke. However, the therapeutic protocols used in the individual studies showed significant variability.

Objective:

A secondary meta-analysis was performed to detect which parameters of these protocols may influence the effect of MT for upper limb paresis after stroke.

Methods:

Trials included in the Cochrane review, which published data for motor function / impairment of the upper limb, were subjected to this analysis. Trials or trial arms that used MT as group therapy or combined it with electrical or magnetic stimulation were excluded. The analysis focused on the parameters mirror size, uni- or bilateral movement execution, and type of exercise. Data were pooled by calculating the total weighted standardized mean difference and the 95% confidence interval.

Results:

Overall, 32 trials were included. The use of a large mirror compared to a small mirror showed a higher effect on motor function. Movements executed unilaterally showed a higher effect on motor function than a bilateral execution. MT exercises including manipulation of objects showed a minor effect on motor function compared to movements excluding the manipulation of objects. None of the subgroup differences reached statistical significance.

Conclusions:

The results of this analysis suggest that the effects on both motor function and impairment of the affected upper limb depend on the therapy protocol. They furthermore indicate that a large mirror, unilateral movement execution and exercises without objects may be parameters that enhance the effects of MT for improving motor function after stroke.

Keywords

Stroke paresis rehabilitation mirror therapy meta-analysis

1 Introduction

Over the last 20 years, MT for treating hemiparesis after stroke has become increasingly popular (Ramachandran & Altschuler, 2009). During mirror therapy (MT), a mirror is placed in the patient’s mid-sagittal plane in such a way that the mirror image of the unaffected limb looks as if it were the affected one (Fig. 1). Several high quality randomized controlled trials in stroke patients have been performed and published, providing good evidence for the beneficial effect of MT - not only on motor function, but also on motor impairment. A recently published systematic review by Thieme and colleagues identified a total number of 62 randomized controlled trials and randomized cross-over trials with 1982 patients (Thieme et al., 2018).

Fig.1

Execution of mirror therapy with realistic perspective caused by a large mirror, while performing body positions unilaterally without the manipulation of objects.

In spite of this bulk of evidence for the effectiveness of MT, however, the precise procedure or therapy protocol is less clear (Bai et al., 2019; Kim et al., 2017; McCabe, 2011). For people suffering from hemiparesis after stroke, there are only few standardized MT protocols, which also differ from each other (Bieniok et al., 2011; Morkisch & Dohle, 2015; Rothgangel & Braun, 2013). The common basis of all therapy studies is the use of a mirror positioned as described above, though there are different sizes of mirrors, different mounts or boxes and sometimes even covers for the unaffected arm (Cacchio et al., 2009a; Hiragami et al., 2012; Michielsen et al., 2011; Mirela et al., 2015). The frequency and duration of the therapy as well as the tasks of the (un-) affected arm that have to be performed during therapy, differ between the researchers’ individual protocols (McCabe, 2011; Selles et al., 2014; Thieme et al., 2018). In particular, there are three major parameters that have to be distinguished in the application of mirror therapy:

Size of the used mirror (full size mirror or “mirror box”-like setup),

Movement execution (unilateral or bilateral), and

Type of exercise (representation of body positions or manipulation of objects, or both in combination).

These parameters occurred in several trials in various combinations (e.g. Lee et al., 2012; Wang et al., 2015). So far, to the best knowledge of the authors, only two studies specifically compared different application variants of MT in patients after stroke (Bai et al., 2019; Selles et al., 2014). Bai and colleagues conducted a randomized controlled trial to compare the effects of MT with and without object manipulation, in patients with mild to moderate upper limb hemiparesis due to subacute stroke. All of these patients performed the movements or tasks bilaterally. The study group that did not manipulate objects, but exclusively represented body positions within the MT (e.g. wrist extension and flexion), showed a better effect on motor impairment measured with the Fugl-Meyer-Assessment (Bai et al., 2019; Fugl-Meyer et al., 1975). In a trial of Selles and colleagues, the effect of different types of movement execution within MT on movement time of a reaching task was investigated. In patients with chronic post-stroke upper limb paresis (Brunnstrom stages III and IV) movement time increases more when movement execution was performed unilaterally rather than bilaterally (Sawner & LaVigne, 1992; Selles et al., 2014). To our knowledge, no trial has investigated the effects of different mirror sizes during MT. Based on the recently published systematic review (Thieme et al., 2018), the aim of this quantitative analysis was to compare different parameters in therapy protocols in order to provide an evidence-based recommendation for the choice of three main points in the application of MT for upper limb hemiparesis after stroke: mirror size, uni- or bilateral movement execution, and type of exercise.

2 Methods

This analysis was a subgroup analysis of the 2018 published update of the Cochrane review “Mirror therapy for improving motor function after stroke” (Thieme et al., 2018). In addition to the original publication, the checklist in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement was applied (Moher et al., 2009).

2.1 Data sources

All information sources for the search and the search algorithm are detailed in the original publication, whose searching process was completed in August of 2017. In that update, 62 randomized controlled trials and randomized cross-over trials were found and included in the qualitative synthesis and the risk of bias in individual trials judgement. 51 randomized controlled trials and randomized cross-over trials were included in the quantitative synthesis (Thieme et al., 2018).

2.2 Study selection

These 51 trials formed the basis for this secondary meta-analysis with a special focus on different parameters for the practical application of MT for paresis of the upper limb.

The inclusion criteria were as follows:

Trials that provided data pertaining to motor function and / or motor impairment of the affected upper limb

Trials that published information regarding the therapy protocol, at least one of the following attributes:

Size of the mirror

Unilateral or bilateral movement execution

Type of exercise (representation of body positions and / or manipulation of objects)

All trials that combined MT with any type of electrical or magnetic stimulation or executed MT as group intervention or used Virtual / Augmented Reality were excluded. If this type of intervention concerned only one of several trial arms, data of this arm were excluded and the remaining at least two trial arms were analyzed (see the PRISMA flow chart, Fig. 2) (Moher et al., 2009).

Fig.2

PRISMA Flow Chart of the selection process.

The primary outcome was motor function at the end of the intervention phase. Motor impairment was defined as a secondary outcome (Thieme et al., 2018).

2.3 Data extraction

Intervention data, characteristics of participants of the qualitative synthesis and outcome data were used from the quantitative synthesis of Thieme and colleagues 2018 which was based on independent assessment of two raters (HT and NM). Additionally, for this review, two raters (NM and KJ) extracted and agreed upon protocol parameters if they were not collected in the Cochrane review process (Thieme et al., 2018).

The three therapy parameters were analyzed as following:

2.3.1 Mirror size

In the first subgroup analysis, the trials were allocated to two groups: use of a large or small mirror. For size classification, the definition by Kim and colleagues were used (Kim et al., 2017). A size of 50×40 cm (to reach eye level) as the minimum for a large mirror was defined. Each smaller height was defined as a small mirror.

Classification was primarily based on the specification of the size of the mirror in the text of the article. In case this information was not available, the image of the set-up was used for classification (if available). In these cases, two authors independently classified the available images. If both pieces of information were missing or unclear, the trial was not considered in this analysis. Studies with special set-ups to cover the unaffected hand were excluded from that analysis.

2.3.2 Uni- or bilateral movement execution

For the second analysis, the trials were allocated to two groups: unilateral or bilateral movement execution as provided in unambiguous terms or as a description in the text. If movements were exclusively executed by the unaffected upper limb, they were classified as a unilateral movement. The classification bilateral movements was used if both the unaffected and the affected upper limbs executed movements during MT.

Trials that used any other form of movement execution, e.g. passive movements of the affected arm by the therapist, were not allocated to either of the two groups.

2.3.3 Type of exercise

For the third analysis, movement instructions regarding excluding or including the manipulation of objects were used as contrast. Representation of body positions (e.g. wrist extension) without using objects were classified as movements excluding the manipulation of objects.

If there was a combination of the representation of body positions with manipulation of objects (e.g. cups), this trial was classified as movements including the manipulation of objects.

2.4 Data analysis

The Review Manager (Review Manager (RevMan), 2014) (Version 5.3) was used for analysis of the continuous data with a random effect model for motor function and motor impairment at the end of the intervention phase. The calculated standardized mean difference (SMD) with 95% confidence intervals (CI) of each trial were pooled through calculation of the overall SMD and 95% CI. The three subgroup analyses were performed for each outcome, motor function and motor impairment, separately.

3 Results

Outcome data of 51 randomized controlled trials or randomized cross-over trials were analyzed in the meta-analysis of Thieme and colleagues (Thieme et al., 2018). Six of these trials focused on lower extremity, and thus were not considered (Arya et al., 2017; In et al., 2016; Kumar, 2013; Marquez et al., 2012; Mohan et al., 2013; Sütbeyaz et al., 2007).

Of the remaining 44 randomized controlled trials and one randomized cross-over trial (Tezuka et al., 2006) that focused on the upper limb, two trials (Moustapha & Rousseaux, 2012; Pandian et al., 2014) provided no data for any of the motor outcomes. Ten trials (Cha & Kim, 2015; Cho & Cha, 2015; Kim et al., 2014; Kim & Lee, 2015; Kojima et al., 2014; Lee et al., 2016; Nagapattinam, 2015; Schick et al., 2017; Yun et al., 2011) were excluded because the therapy protocol contained a combination of MT with any type of electrical or magnetic stimulation. One trial employed a Virtual Reality set up (In et al., 2012). For five trials, one of the trial arms had to be excluded due to the use of any type of electrical or magnetic stimulation in four trials and the use of MT as a group intervention in one trial (Amasyali & Yaliman, 2016; Ji et al., 2014; Lin et al., 2014; Thieme et al., 2013; Wang et al., 2015).

Finally, a total number of 31 randomized controlled trials and one randomized cross-over trial with 1031 participants fulfilled the inclusion criteria and were included in the quantitative analysis, even though information about all three aspects was not available for every trial (Table 1, Fig. 2).

Table 1
Overview of trials‘ therapy protocols

ID of the analyzed trial Outcome Mirror size Uni- or bilateral movement execution Type of exercise

Motor function Motor impairment Large mirror Small mirror Unilateral execution Bilateral execution Movements excluding manipulation of objects Movements including manipulation of objects

Acerra 2007 • • • – – • – •

Alibakshi 2016 • • –¹ –¹ – • – –

Amasyali 2016 • • – • • – • –

Arya 2015 – • • – • – – •

Bae 2012 • – – – – • • –

Cacchio 2009a • – • – • – • –

Cacchio 2009b • – –¹ –¹ • – • –

Colomer 2016 • • –¹ –¹ • – – •

Dohle 2009 • • • – – • • –

Gurbuz 2016 – • – – • – • –

Hiragami 2012 • • – • – • – •

Invernizzi 2013 • • • – • – • –

Ji 2014 • • – • • – • –

Kim 2016 • • • – • – – •

Kuzgun 2012 – • • – – – – –

Lee 2012 • • – • – • • –

Lim 2016 – • –² –² – • – •

Lin 2014 • • • – – • – •

Michielsen 2011 • • –² –² – • – •

Mirela 2015 – • – • – • • –

Park 2015 • – –¹ –¹ • – – •

Piravej 2012 • • –¹ –¹ – – – •

Rodriguez 2016 • • – • – • – •

Samuelkamaleshkumar 2014 • • – • • – – •

Seok 2010 • • – • • – • –

Tezuka 2006 – • • – –² –² – •

Thieme 2013 • • • – – • – •

Tyson 2015 • • – • – • – •

Wang 2015 • • • – • – – •

Wu 2013 – • • – – • – •

Yavuzer 2008 – • –² –² – • • –

Yoon 2014 • • • – • – – •

Total n = 32 24 28 13 9 14 15 12 18

ID of the analyzed trial	Outcome	Mirror size	Uni- or bilateral movement execution	Type of exercise
	Motor function	Motor impairment	Large mirror	Small mirror	Unilateral execution	Bilateral execution	Movements excluding manipulation of objects	Movements including manipulation of objects
Acerra 2007	•	•	•	–	–	•	–	•
Alibakshi 2016	•	•	–¹	–¹	–	•	–	–
Amasyali 2016	•	•	–	•	•	–	•	–
Arya 2015	–	•	•	–	•	–	–	•
Bae 2012	•	–	–	–	–	•	•	–
Cacchio 2009a	•	–	•	–	•	–	•	–
Cacchio 2009b	•	–	–¹	–¹	•	–	•	–
Colomer 2016	•	•	–¹	–¹	•	–	–	•
Dohle 2009	•	•	•	–	–	•	•	–
Gurbuz 2016	–	•	–	–	•	–	•	–
Hiragami 2012	•	•	–	•	–	•	–	•
Invernizzi 2013	•	•	•	–	•	–	•	–
Ji 2014	•	•	–	•	•	–	•	–
Kim 2016	•	•	•	–	•	–	–	•
Kuzgun 2012	–	•	•	–	–	–	–	–
Lee 2012	•	•	–	•	–	•	•	–
Lim 2016	–	•	–²	–²	–	•	–	•
Lin 2014	•	•	•	–	–	•	–	•
Michielsen 2011	•	•	–²	–²	–	•	–	•
Mirela 2015	–	•	–	•	–	•	•	–
Park 2015	•	–	–¹	–¹	•	–	–	•
Piravej 2012	•	•	–¹	–¹	–	–	–	•
Rodriguez 2016	•	•	–	•	–	•	–	•
Samuelkamaleshkumar 2014	•	•	–	•	•	–	–	•
Seok 2010	•	•	–	•	•	–	•	–
Tezuka 2006	–	•	•	–	–²	–²	–	•
Thieme 2013	•	•	•	–	–	•	–	•
Tyson 2015	•	•	–	•	–	•	–	•
Wang 2015	•	•	•	–	•	–	–	•
Wu 2013	–	•	•	–	–	•	–	•
Yavuzer 2008	–	•	–²	–²	–	•	•	–
Yoon 2014	•	•	•	–	•	–	–	•
Total n = 32	24	28	13	9	14	15	12	18

¹Provided no information on the mirrors’ size; ²Could not be included into this analysis due to different design of the set up or the type of execution.

All three subgroup analyses could only be performed for 20 trials (62.5 %). Data on motor impairment for the trial of Hiragami and colleagues, that could not be included in the Cochrane review of Thieme and colleagues, could be included in this meta-analysis because the contact with the authors was resumed (Hiragami et al., 2012; Thieme et al., 2018). The mean age of the participants (58.4 years) ranged from 46 years (Arya et al., 2015) to 68 years (Acerra, 2007). There were more males (60%) than females, more participants with an ischemic stroke (70%) than a hemorrhagic stroke, and more participants with a left-sided paresis (53%) than a right-sided paresis of the upper limb included in those trials. The sample size ranged from 14 (Hiragami et al., 2012) to 94 (Tyson et al., 2015) participants.

3.1 Effects of mirror size

The size of the mirror was the least published information (78 %), but showed the greatest variance in the design, for example: different self-made constructions, or stool-like covers to prevent looking at the unaffected arm (Cacchio et al., 2009a; Hiragami et al., 2012; Michielsen et al., 2011; Mirela et al., 2015).

Tezuka and colleagues provided no data on the mirror size, but cited the trial of Sütbeyaz and colleagues (not included, see results) which published the size of the mirror in the text of the article (Sütbeyaz et al., 2007; Tezuka et al., 2006). Thus, the trial of Tezuka and colleagues was included with these data (Tezuka et al., 2006). Two trials provided no information about the set-up (Bae et al., 2012; Gurbuz et al., 2016). Five trials published insufficient information, therefore a classification was not possible (Alibakhshi et al., 2016; Cacchio et al., 2009b; Colomer et al., 2016; Park et al., 2015; Piravej et al., 2012). Three trials used a cover over the unaffected hand and were excluded from this analysis (Lim et al., 2016; Michielsen et al., 2011; Yavuzer et al., 2008).

Thus, a total of 22 trials could be included in this analysis - 17 trials for the outcome motor function and 21 trials for the outcome motor impairment (Table 1).

3.1.1 Effect on motor function

In this analysis, the experimental groups (EXP) of nine trials (317 participants: 156 EXP, 161 CON (control group)) (Acerra, 2007; Cacchio et al., 2009a; Dohle et al., 2009; Invernizzi et al., 2013; Kim et al., 2016; Lin et al., 2014; Thieme et al., 2013; Wang et al., 2015; Yoon et al., 2014) used a large mirror and the experimental groups of eight trials (241 participants: 134 EXP, 107 CON) (Amasyali & Yaliman, 2016; Hiragami et al., 2012; Ji et al., 2014; Lee et al., 2012; Rodrigues et al., 2016; Samuelkamaleshkumar et al., 2014; Seok et al., 2010; Tyson et al., 2015) used a small mirror. MT performed with a large mirror had a statistically significant effect on motor function (SMD 0.77, 95% CI 0.20 to 1.33; I² = 82 %) in participants with paresis of the upper limb following stroke. A considerably smaller, but nevertheless statistically significant effect on motor function was found for the use of a small mirror (SMD 0.28, 95% CI 0.02 to 0.54; I² = 0%) (Fig. 3). The difference between subgroups was statistically non-significant (P = 0.12).

Fig.3

Forest plot of all trials that used a large mirror versus all trials that used a small mirror (Outcome: Motor function).

3.1.2 Effect on motor impairment

The experimental groups of 12 trials (372 participants: 185 EXP, 187 CON) (Acerra, 2007; Arya et al., 2015; Dohle et al., 2009; Invernizzi et al., 2013; Kim et al., 2016; Kuzgun et al., 2012; Lin et al., 2014; Tezuka et al., 2006; Thieme et al., 2013; Wang et al., 2015; Wu et al., 2013; Yoon et al., 2014) used a large mirror. The analysis showed a statistically significant effect on motor impairment in participants with upper limb paresis after stroke (SMD 0.62, 95% CI 0.27 to 0.98; I² = 62%). In the analysis of using a small mirror within the MT, nine trials with a total of 256 participants (141 EXP, 115 CON) (Amasyali & Yaliman, 2016; Hiragami et al., 2012; Ji et al., 2014; Lee et al., 2012; Mirela et al., 2015; Rodrigues et al., 2016; Samuelkamaleshkumar et al., 2014; Seok et al., 2010; Tyson et al., 2015) were included. There was a statistically non-significant effect on motor impairment for this type of device (SMD 0.26, 95% CI –0.06 to 0.57; I² = 28%). Subgroup differences did not demonstrate statistical significance (P = 0.13).

3.2 Effects of uni- or bilateral movement execution

Two trials provided no information on a uni- or bilateral upper limb movement execution (Kuzgun et al., 2012; Piravej et al., 2012). The trial of Tezuka and co-workers was not included, because the therapist passively moved the affected side (Tezuka et al., 2006).

In summary, a total of 29 trials could be included in this subgroup analysis- 23 trials for the outcome of motor function and 25 for the outcome of motor impairment (Table 1).

3.2.1 Effect on motor function

In 12 trials with a total of 360 participants (174 EXP, 186 CON) (Amasyali & Yaliman, 2016; Cacchio et al., 2009a, 2009b; Colomer et al., 2016; Invernizzi et al., 2013; Ji et al., 2014; Kim et al., 2016; Park et al., 2015; Samuelkamaleshkumar et al., 2014; Seok et al., 2010; Wang et al., 2015; Yoon et al., 2014) unilateral movements were executed within MT. The analysis showed a statistically significant effect on motor function (SMD 0.69, 95% CI 0.11 to 1.27; I² = 84%). In the analysis of trials executing bilateral movements within MT, 11 trials with a total of 367 participants (196 EXP, 171 CON) (Acerra, 2007; Alibakhshi et al., 2016; Bae et al., 2012; Dohle et al., 2009; Hiragami et al., 2012; Lee et al., 2012; Lin et al., 2014; Michielsen et al., 2011; Rodrigues et al., 2016; Thieme et al., 2013; Tyson et al., 2015) were included. A considerably smaller, but nevertheless statistically significant effect on motor function was also found for the bilateral execution of movements within the MT (SMD 0.36, 95% CI 0.14 to 0.59; I² = 12%) (Fig. 4). The test for subgroup differences did not reach statistical significance (P = 0.31).

Fig.4

Forest plot of all trials that used unilateral movement execution versus all trials that used bilateral movement execution (Outcome: Motor function).

3.2.2 Effect on motor impairment

Data for motor impairment were presented in 11 trials with a total of 322 participants (160 EXP, 162 CON) (Amasyali & Yaliman, 2016; Arya et al., 2015; Colomer et al., 2016; Gurbuz et al., 2016; Invernizzi et al., 2013; Ji et al., 2014; Kim et al., 2016; Samuelkamaleshkumar et al., 2014; Seok et al., 2010; Wang et al., 2015; Yoon et al., 2014) that executed unilateral movements within the MT. This analysis showed a statistically significant effect on motor impairment in participants with paresis of the upper limb after stroke (SMD 0.56, 95% CI 0.10 to 1.03; I² = 75%). In 14 trials with a total of 493 participants (257 EXP, 236 CON) (Acerra, 2007; Alibakhshi et al., 2016; Dohle et al., 2009; Hiragami et al., 2012; Lee et al., 2012; Lim et al., 2016; Lin et al., 2014; Michielsen et al., 2011; Mirela et al., 2015; Rodrigues et al., 2016; Thieme et al., 2013; Tyson et al., 2015; Wu et al., 2013; Yavuzer et al., 2008) bilateral movements were executed within the MT. The analysis showed also a statistically significant effect (SMD 0.40; 95 % CI 0.15 to 0.64; I² = 40%). The subgroup differences did not demonstrate statistical significance (P = 0.53).

3.3 Effects of type of exercise

Two of trials provided no information on type of exercise, therefore they were excluded from this analysis (Alibakhshi et al., 2016; Kuzgun et al., 2012).

A total of 30 trials presented information about the type of exercise - 23 trials for the outcome motor function and 27 for the outcome motor impairment (Table 1).

3.3.1 Effect on motor function

In ten trials the participants (276 participants: 133 EXP, 143 CON) (Amasyali & Yaliman, 2016; Bae et al., 2012; Cacchio et al., 2009a, 2009b; Dohle et al., 2009; Invernizzi et al., 2013; Ji et al., 2014; Lee et al., 2012; Seok et al., 2010; Yoon et al., 2014) were asked to represent body positions, excluding the manipulation of objects within MT. In thirteen trials with a total of 460 participants (242 EXP, 218 CON) (Acerra, 2007; Colomer et al., 2016; Hiragami et al., 2012; Kim et al., 2016; Lin et al., 2014; Michielsen et al., 2011; Park et al., 2015; Piravej et al., 2012; Rodrigues et al., 2016; Samuelkamaleshkumar et al., 2014; Thieme et al., 2013; Tyson et al., 2015; Wang et al., 2015), the execution of both exercise types was required within MT, representation of body positions and manipulation of objects. The subgroup analysis presented a statistically significant effect on motor function (SMD 0.67, 95% CI 0.18 to 1.16; I² = 73 %), when the manipulation of objects was excluded within MT. There was a statistically non-significant effect on motor function, when the required movements contained the manipulation of objects (SMD 0.39, 95 % CI –0.03 to 0.80; I² = 77 %) (Fig. 5). However, subgroup differences did not demonstrate statistical significance (P = 0.39).

Fig.5

Forest plot of all trials that excluded manipulation of objects versus all trials that included manipulation of objects (Outcome: Motor function).

3.3.2 Effect on motor impairment

The experimental groups of 11 trials (286 participants: 141 EXP, 145 CON) (Amasyali & Yaliman, 2016; Dohle et al., 2009; Gurbuz et al., 2016; Invernizzi et al., 2013; Ji et al., 2014; Kuzgun et al., 2012; Lee et al., 2012; Mirela et al., 2015; Seok et al., 2010; Yavuzer et al., 2008; Yoon et al., 2014) excluded object manipulation while performing the movements within MT. The analysis showed a statistically significant effect on motor impairment in participants with upper limb paresis after stroke (SMD 0.42, 95% CI 0.18 to 0.67; I² = 7 %).

In the analysis of a combination of object manipulation and representation of body positions within the MT, 16 trials with a total of 573 participants (300 EXP, 273 CON) (Acerra, 2007; Arya et al., 2015; Colomer et al., 2016; Hiragami et al., 2012; Kim et al., 2016; Lim et al., 2016; Lin et al., 2014; Michielsen et al., 2011; Piravej et al., 2012; Rodrigues et al., 2016; Samuelkamaleshkumar et al., 2014; Tezuka et al., 2006; Thieme et al., 2013; Tyson et al., 2015; Wang et al., 2015; Wu et al., 2013) were included. There was also a statistically significant effect on motor impairment for this type of exercise (SMD 0.43, 95 % CI 0.10 to 0.75; I² = 70%). Between subgroups, there was a statistically non-significant difference (P = 0.99).

4 Discussion

The aim of this meta-analysis was to gain knowledge about the effect of different parameters in MT protocols for upper limb paresis after stroke. The results suggest that both motor function and motor impairment of the affected upper limb depend on the therapy protocol. Despite the lack of significant differences between subgroups, the results indicate that a large mirror, unilateral movement execution and exercises without objects lead to better results regarding motor function and motor impairment.

4.1 Effects of mirror size

Regarding the outcome parameters, motor function as well as motor impairment, the effects of a large mirror were greater than the effects of a small mirror within the included trials.

The mirror illusion is produced during the observation of the moving reflected limb in the mirror (Fig. 1). A large mirror enables a realistic perspective to be formed. That means it supports the perception that the mirror image complements the body of the person. Results of studies in the field of Virtual Reality showed the importance of body presentation design and its influence on presence and body ownership (Gonzalez-Franco et al., 2010; Sanchez-Vives & Slater, 2005; Waltemate et al., 2018).

The benefit of having a large mirror might be that it allows a less distracted unilateral visual feedback. That means the observation of the mirror image is less distracted by the moving unaffected arm, whereas a small mirror tends to produce bilateral visual feedback. From a therapeutic point of view, it is important to adapt the physical and cognitive demands of the therapy content to the patients. A large mirror offers another advantage in that case. As proximal parts of the arm could also be included, an adaptation of the content of therapy, called Shaping, is possible (Hömberg, 2013; Morkisch et al., 2017a, 2017b). For example, the standardized therapy protocol developed by Morkisch and Dohle provided up to 125 potential combinations of distal and proximal parts of the arm for adaption of the therapy content. The number of combinations is selected on the basis on the patients’ level of performance. That shaping mechanism could optimize the patients’ visual attention to observe the mirror illusion (Bieniok et al., 2011; Morkisch & Dohle, 2015; Morkisch et al., 2017b). Using a small mirror does not provide this opportunity: In that case, only a very low number of different tasks can be required, which bears the risk of introducing repetitive tasks that do not require constant attention towards the mirror image.

4.2 Effects of uni- or bilateral movement execution

In our analysis, a unilateral movement execution showed a greater effect on both motor function and motor impairment than a bilateral movement execution. This is in accordance with a study of Selles and co-workers that directly compared different therapy variants in patients with an upper limb paresis after stroke. In their analysis of improvement of movement time, unilaterally executed MT did not statistically differ from a training of the affected arm only. However, bilaterally executed MT was less effective. The authors assumed greater fatigue of the affected arm in those conditions (Selles et al., 2014).

Furthermore, bilateral execution requires split attention to both body sides and thus may reduce attention to the affected arm (and its image). This consideration is also true for conventional bilateral training (without a mirror), as indicated in three recent systematic reviews (Hatem et al., 2016; Pollock et al., 2014; Veerbeek et al., 2014). However, for MT, there might be a selection effect as well. “True” bilateral performance of MT is only possible for mild paresis when active movements of the affected limb are possible at all. In this context, McCabe and colleagues already showed in 2011 that not exactly identical movements of the arm in front of and behind the mirror can lead to unpleasant sensations and even pain, possibly further degrading the therapy effect (McCabe, 2011).

The advantage of the mirror illusion is to influence the brain’s networks without moving the affected upper limb at the same time. Several imaging studies showed a sufficient neural activation in the observation of unilateral self-performed, mirrored movements without movements of the affected limb at all (Dohle et al., 2004; Mehnert et al., 2013; Wang et al., 2013).

4.3 Effects of type of exercise

Regarding motor function, realization of MT without object manipulation showed a greater effect than therapy variants including object manipulation. For the outcome parameter motor impairment, the effects of inclusion or exclusion of object manipulation were similar. This is in line with a recently published randomized controlled trial that compared MT variants with and without use of objects in patients with mild to moderate hemiparesis. The authors reported a tendency for greater improvement in motor impairment for MT without the manipulation of objects (Bai et al., 2019).

This view was recently challenged by an fMRI study of Manuweera and co-workers, reporting greater activation of the superior parietal lobe extending into the intraparietal sulcus and its coupling with the primary motor cortex in mirrored conditions with object (target) compared to conditions without an object (target) (Manuweera et al., 2019). It should be considered, however, that this study only reported brain activation measures during performance of the task, but no behavioral data on the effect of mirror training under these conditions. Up to now, it is unclear whether this increased brain activation does in fact transform into behavioral improvement or just reflects compensatory activity due to an increase in task complexity. In fact, there were other areas in the primary somatosensory cortex and bilateral insula that are more active in the conditions without a target.

Beyond that, the results of this subgroup analysis seem to contradict common assumptions on motor learning. A recently published systematic review on motor rehabilitation to upper limb paresis after stroke recommended the incorporation of task-oriented approaches into the rehabilitation process (Veerbeek et al., 2014). However, Hatem and colleagues argued that the motor enhancements were not translated into the activities of daily living (Hatem et al., 2016). During MT, a task-oriented training via manipulation of objects is basically possible as well. However, this requires a coordinate transformation in addition to the body transformation already introduced by the mirror. Imaging results showed that body and coordinate transformation are accomplished by distinct and different neural structures (Dohle et al., 2011). Additional coordinate transformation does not only reduce activity in the precunei of either hemisphere, but also movement accuracy. Thus, coordinate transformations in a mirror can be performed and trained, but this might reduce neural activity and attention to the mirror illusion itself.

It might be argued that MT without objects might be too low demanding in order to keep sustained attention to the mirror image. However, the level of demand could be increased in other ways as well. For example, the Bonn and Berlin therapy protocols allowed the combination of proximal and distal body postures, thus varying task difficulty (Bieniok et al., 2011; Morkisch & Dohle, 2015). Additional shaping items such as instruction pace could further modify the level of demand (Morkisch et al., 2017a).

4.4 Limitations of the study

None of the differences in the three analyses was significant. This might also be due to the fact that the majority of studies was included in more than one of the analyses. Thus, possible differences between the clinical effects could not be unambiguously attributed to only one factor. Nevertheless, the direction of the differences provided strong suggestions on the “optimum” therapy protocol, which is in accordance with neurophysiological mechanisms (as detailed above).

Despite the application of a random effect model, the heterogeneity of the single subgroup analyses was very high, already in the original and other reviews on MT published previously (Cantero-Téllez et al., 2018; Hartman & Altschuler, 2016; Thieme et al., 2018). In the Cochrane review, subgroup analyses regarding the study population, e.g. stage of stroke, could not reduce this heterogeneity (Thieme et al., 2018).

Some of the trials only included stroke patients with sensory deficits or the presence of CRPS (Cacchio et al., 2009a, 2009b; Colomer et al., 2016). Notably, these trials provided especially outlier results (Fig. 3–5).

4.5 Implications for further research

In general, as already proposed by Hartman and Altschuler, further analyses regarding the MT effect should try to identify subgroups of patients who may derive benefit from MT (Hartman & Altschuler, 2016). In a pilot study, Brunetti and collaborators suggested to use the change of brain activation due to the mirror illusion as a predictor (Brunetti et al., 2015). Another approach using routine data, identified the sensory deficit, especially 2-point discrimination, as a potential predictor (Seifert et al., 2017). However, based on the present analysis, it cannot be excluded that different patients (e.g. hemiparesis, sensory deficit or pain) respond to different therapy protocols differently. To the best knowledge of the authors, this has not yet been investigated in detail.

5 Conclusion

The included trials showed a high variance in the actual execution of MT. The results of this meta-analysis indicate that improvement of both motor function and motor impairment of the affected upper limb after stroke depend on the therapy protocol. The use of a large mirror, a unilateral movement execution and exercises excluding the manipulation of objects were shown to result in greater effects. These are aspects of the therapy protocols proposed by the authors previously (Bieniok et al., 2011; Morkisch & Dohle, 2015). However, the results should be interpreted with caution, because the tests for subgroup differences did not reach statistical significance.

Due to the existing bulk of evidence, further studies aiming to demonstrate the efficacy of MT as an adjunct therapy or compared to a placebo therapy are no longer necessary (Thieme et al., 2018). Rather, new studies should consider specific questions regarding the optimal application of MT after stroke. Ultimately, this should result in the definition of an optimum therapy protocol and an optimum target population for applying MT after stroke.

Conflict of interest statement

HT and CD are each author of an included study. Both authors were neither involved in checking these trisals for eligibility, nor did they extract data from these trials.

Both HT and CD have received and will receive honorarium for presentations and seminars on mirror therapy and have received grants from the Federal Ministry of Education and Research for studies on conventional and technical-assisted mirror therapy.

CD and NM are authors of therapy manuals on mirror therapy, published at Schulz-Kirchner-Verlag (CD) and Hippocampus Verlag (CD and NM).

Footnotes

Acknowledgments

This study was not supported by any grant. Assistance provided by Katrin Jettkowski was greatly appreciated.

References

Acerra,

N.E.

(2007). Is early post-stroke upper limb mirror therapy associated with improved sensation and motor recovery? A randomised-controlled trial [PhD thesis]. Sensorimotor Dysfunction in CRPS1 and Stroke: Characteristics, Prediction and Intervention. University of Queensland, Brisbane, Australia.

Alibakhshi,

, Samaei,

, Khalili,

, & Siminghalam,

(2016). A comparative study on the effects of mirror therapy and bilateral arm training on hand function of chronic hemiparetic patients. Koomesh Journal of Semnan University of Medical Sciences, 17(3), 589–595.

Amasyali,

S.Y

, & Yaliman,

(2016). Comparison of the effects of mirror therapy and electromyography-triggered neuromuscular stimulation on hand functions in stroke patients: A pilot study. International Journal of Rehabilitation Research. Internationale Zeitschrift Fur Rehabilitationsforschung. . Revue Internationale De Recherches De Readaptation, 39(4), 302–307. https://doi.org/10.1097/MRR.0000000000000186

Arya,

K.N

, Pandian,

, Kumar,

, & Puri,

(2015). Task-based mirror therapy augmenting motor recovery in poststroke hemiparesis: A randomized controlled trial. Journal of Stroke and Cerebrovascular Diseases, 24(8), 1738–1748. https://doi.org/10.1016/j.jstrokecerebrovasdis.2015.03.026

Arya,

K.N

, Pandian,

, & Kumar,

(2017). Effect of activity-based mirror therapy on lower limb motor-recovery and gait in stroke: A randomised controlled trial. Neuropsychological Rehabilitation 1–18. https://doi.org/10.1080/09602011.2017.1377087

Bae,

, Jeong,

, & Kim,

(2012). Effects of mirror therapy on subacute stroke patients’ brain waves and upper extremity functions. Journal of Physical Therapy Science, 24(11), 1119–1122. https://doi.org/10.1589/jpts.24.1119

Bai,

, Zhang,

, Shu,

, & Niu,

(2019). Comparison between movement-based and task-based mirror therapies on improving upper limb functions in patients with stroke: A pilot randomized controlled trial. Frontiers in Neurology 10–https://doi.org/10.3389/fneur.2019.00288

Bieniok,

, Govers,

, & Dohle,

(2011) Spiegeltherapie in der Neurorehabilitation (2. Aufl). Idstein: Schulz-Kirchner.

Brunetti,

, Morkisch,

, Fritzsch,

, Mehnert,

, Steinbrink,

, Niedeggen,

, & Dohle,

(2015). Potential determinants of efficacy of mirror therapy in stroke patients–A pilot study. Restorative Neurology and Neuroscience, 33(4), 421–434. https://doi.org/10.3233/RNN-140421

10.

Cacchio,

, De Blasis,

, Santilli,

, & Spacca,

, (2009a). Mirror therapy in complex regional pain syndrome type 1 of the upper limb in stroke patients. Neurorehabilitation and Neural Repair, 23(8), 792–799. https://doi.org/10.1177/1545968309335977

11.

Cacchio,

, De Blasis,

, Necozione,

, di Orio,

, & Santilli,

, (2009b). Mirror therapy for chronic complex regional pain syndrome type 1 and stroke. New England Journal of Medicine, 361(6), 634–636. https://doi.org/10.1056/NEJMc0902799

12.

Cantero-Téllez,

, Naughton,

, Algar,

, & Valdes,

(2018). Outcome measurement of hand function following mirror therapy for stroke rehabilitation: A systematic review.. Journal of Hand Therapyhttps://doi.org/10.1016/j.jht.2018.01.009

13.

Cha,

H.G

, & Kim,

M.-K.

(2015). Therapeutic efficacy of low frequency transcranial magnetic stimulation in conjunction with mirror therapy for sub-acute stroke patients. Journal of Magnetics, 20(1), 52–56. https://doi.org/10.4283/JMAG.2015.20.1.052

14.

Cho,

H.-S

, & Cha,

H.-G.

(2015). Effect of mirror therapy with tDCS on functional recovery of the upper extremity of stroke patients. Journal of Physical Therapy Science, 27(4), 1045–1047. https://doi.org/10.1589/jpts.27.1045

15.

Colomer,

, Noé,

, & Llorens,

(2016). Mirror therapy in chronic stroke survivors with severely impaired upper limb function: A randomized controlled trial. European Journal of Physical and Rehabilitation Medicine, 52(3), 8.

16.

Dohle,

, Kleiser,

, Seitz,

R.J

, & Freund,

H.-J.

(2004). Body scheme gates visual processing. Journal of Neurophysiology, 91(5), 2376–2379. https://doi.org/10.1152/jn.00929.2003

17.

Dohle,

, Püllen,

, Nakaten,

, Küst,

, Rietz,

, & Karbe,

(2009). Mirror therapy promotes recovery from severe hemiparesis: A randomized controlled trial. Neurorehabilitation and Neural Repair, 23(3), 209–217. https://doi.org/10.1177/1545968308324786

18.

Dohle,

, Stephan,

K.M

, Valvoda,

J.T

, Hosseiny,

, Tellmann,

, Kuhlen,

, & Freund,

H.-J.

(2011). Representation of virtual arm movements in precuneus. Experimental Brain Research, 208(4), 543–555. https://doi.org/10.1007/s00221-010-2503-0

19.

Fugl-Meyer,

A.R

, Jääskö,

, Leyman,

, Olsson,

, & Steglind,

(1975). The post-stroke hemiplegic patient. 1. a method for evaluation of physical performance. Scandinavian Journal of Rehabilitation Medicine, 7(1), 13–31.

20.

Gonzalez-Franco,

, Perez-Marcos,

, Spanlang,

, & Slater,

(2010). The contribution of real-time mirror reflections of motor actions on virtual body ownership in an immersive virtual environment. 2010 IEEE Virtual Reality Conference (VR) 111–114. https://doi.org/10.1109/VR.2010.5444805

21.

Gurbuz,

, Afsar,

S.I

, Ayaş,

, & Cosar,

S.N.

(2016). Effect of mirror therapy on upper extremity motor function in stroke patients: A randomized controlled trial. Journal of Physical Therapy Science, 28(9), 2501–2506. https://doi.org/10.1589/jpts.28.2501

22.

Hartman,

, & Altschuler,

E.L.

(2016). Mirror therapy for hemiparesis following stroke: A review. Current Physical Medicine and Rehabilitation Reports, 4(4), 237–248. https://doi.org/10.1007/s40141-016-0131-8

23.

Hatem,

S.M

, Saussez,

, della Faille

, Prist,

, Zhang,

, Dispa,

, & Bleyenheuft,

(2016). Rehabilitation of motor function after stroke: A multiple systematic review focused on techniques to stimulate upper extremity recovery. Frontiers in Human Neuroscience 10–https://doi.org/10.3389/fnhum.2016.00442

24.

Hiragami,

, Sato,

, Harada,

, & Kagawa,

(2012). The effect of mirror therapy on finger motor dysfunction after stroke. Journal of Japanese Society of Physical Therapy, 39(5), 330–337. https://doi.org/10.15063/rigaku.KJ00008159427

25.

Hömberg,

(2013). Neurorehabilitation approaches to facilitate motor recovery. In Handbook of Clinical Neurology (Vol. 110,) 161–173 https://doi.org/10.1016/B978-0-444-52901-5.00014-9

26.

In,

T.S

, Jung,

K.S

, Lee,

S.W

, & Song,

C.H.

(2012). Virtual reality reflection therapy improves motor recovery and motor function in the upper extremities of people with chronic stroke. Journal of Physical Therapy Science, 24(4), 339–343. https://doi.org/10.1589/jpts.24.339

27.

In,

T.S

, Lee,

, & Song,

(2016). Virtual reality reflection therapy improves balance and gait in patients with chronic stroke: Randomized controlled trials. Medical Science Monitor, 22, 4046–4053 https://doi.org/10.12659/MSM.898157

28.

Invernizzi,

, Negrini,

, Carda,

, Lanzotti,

, Cisari,

, & Baricich,

(2013). The value of adding mirror therapy for upper limb motor recovery of subacute stroke patients: A randomized controlled trial. European Journal of Physical and Rehabilitation Medicine, 49(3), 311–317.

29.

Ji,

S.-G

, Cha,

H.-G

, & Kim,

M.-K.

(2014). Stroke recovery can be enhanced by using repetitive transcranial magnetic stimulation combined with mirror therapy. Journal of Magnetics, 19(1), 28–31. https://doi.org/10.4283/JMAG.2014.19.1.028

30.

Kim,

, Lee,

, & Song,

(2014). Effect of functional electrical stimulation with mirror therapy on upper extremity motor function in poststroke patients. Journal of Stroke and Cerebrovascular Diseases: The Official Journal of National Stroke Association, 23(4), 655–661. https://doi.org/10.1016/j.jstrokecerebrovasdis.2013.06.017

31.

Kim,

J.H

, & Lee,

B.-H.

(2015). Mirror therapy combined with biofeedback functional electrical stimulation for motor recovery of upper extremities after Stroke: A pilot randomized controlled trial: Mirror therapy combined with bio-feedback FES. Occupational Therapy International, 22(2), 51–60 https://doi.org/10.1002/oti.1384

32.

Kim,

, Yi,

, & Song,

C.-H.

(2017). Kinematic analysis of head, trunk, and pelvic motion during mirror therapy for stroke patients. Journal of Physical Therapy Science, 29(10), 1793–1799 https://doi.org/10.1589/jpts.29.1793

33.

Kim,

, Lee,

, Kim,

, Lee,

, & Kim,

(2016). Effects of mirror therapy combined with motor tasks on upper extremity function and activities daily living of stroke patients. Journal of Physical Therapy Science, 28(2), 483–487. https://doi.org/10.1589/jpts.28.483

34.

Kojima,

, Ikuno,

, Morii,

, Tokuhisa,

, Morimoto,

, & Shomoto,

(2014). Feasibility study of a combined treatment of electromyography-triggered neuromuscular stimulation and mirror therapy in stroke patients: A randomized crossover trial. NeuroRehabilitation, 34(2), 235–244. https://doi.org/10.3233/NRE-131038

35.

Kumar,

(2013). Effect of functional electrical stimulation with mirror therapy on upper extremity motor function in poststroke patients [PhD thesis] (PhD thesis). Rajiv Gandhi University of Health Sciences, Karnataka, Bangalore, India.

36.

Kuzgun,

, Ozgen,

, Armagan,

, Tascioglu,

, & Baydemir,

(2012). The efficacy of mirror therapy combined with conventional stroke rehabilitation program on motor and functional recovery. Turkish Journal of Cerebrovascular Diseases, 18(3), 77–82. https://doi.org/10.5505/tbdhd.2012.14622

37.

Lee,

, Lee,

, & Jeong,

(2016). Mirror therapy with neuromuscular electrical stimulation for improving motor function of stroke survivors: A pilot randomized clinical study. Technology and Health Care, 24(4), 503–511. https://doi.org/10.3233/THC-161144

38.

Lee,

M.M

, Cho,

, & Song,

C.H.

(2012). The mirror therapy program enhances upper-limb motor recovery and motor function in acute stroke patients: American Journal of Physical Medicine & Rehabilitation. 91(8), 689–700. https://doi.org/10.1097/PHM.0b013e31824fa86d

39.

Lim,

K.-B

, Lee,

H.-J

, Yoo,

, Yun,

H.-J

, & Hwang,

H.-J.

(2016). Efficacy of mirror therapy containing functional tasks in poststroke patients. Annals of Rehabilitation Medicine, 40(4), 629 https://doi.org/10.5535/arm.2016.40.4.629

40.

Lin,

, Huang,

, Chen,

, Wu,

, & Huang,

(2014). Combining afferent stimulation and mirror therapy for rehabilitating motor function, motor control, ambulation, and daily functions after stroke. Neurorehabilitation and Neural Repair, 28(2), 153–162. https://doi.org/10.1177/1545968313508468

41.

Manuweera,

, Yarossi,

, Adamovich,

, & Tunik,

(2019). Parietal activation associated with target-directed right hand movement is lateralized by mirror feedback to the ipsilateral hemisphere. Frontiers in Human Neuroscience 12. https://doi.org/10.3389/fnhum.2018.00531

42.

Marquez,

, Hollingsworth,

, & Lancaster,

(2012). It’s all just stroke and mirrors! The clinical implementation of mirror therapy to restore lower limb function and mobility following stroke and traumatic brain injury. Cerebrovascular Diseases, 33(Suppl 2), 843.

43.

McCabe,

(2011). Mirror visual feedback therapy. A practical approach. Journal of Hand Therapy, 24(2), 170–179. https://doi.org/10.1016/j.jht.2010.08.003

44.

Mehnert,

, Brunetti,

, Steinbrink,

, Niedeggen,

, & Dohle,

(2013). Effect of a mirror-like illusion on activation in the precuneus assessed with functional near-infrared spectroscopy. Journal of Biomedical Optics, 18(6), 066001 https://doi.org/10.1117/1.JBO.18.6.066001

45.

Michielsen,

M.E

, Selles,

R.W

, Ribbers,

G.M

, & Bussmann,

J. B. J.

(2011). Motor recovery and cortical reorganization after mirror therapy in chronic stroke patients: A phase II randomized controlled trial. Neurorehabilitation and Neural Repair 12 https://https://doi.org/10.1177/1545968310385127

46.

Mirela,

C.L

, Matei,

, Ignat,

, & Popescu,

C.D.

(2015). Mirror therapy enhances upper extremity motor recovery in stroke patients. Acta Neurologica Belgica, 115(4), 597–603. https://doi.org/10.1007/s13760-015-0465-5

47.

Mohan,

, Babu,

, Kumar,

, Suresh,

, Misri,

, & Chakrapani,

(2013). Effectiveness of mirror therapy on lower extremity motor recovery, balance and mobility in patients with acute stroke: A randomized sham-controlled pilot trial. Annals of Indian Academy of Neurology, 16(4), 634–639. https://doi.org/10.4103/0972-2327.120496

48.

Moher,

, Liberati,

, Tetzlaff,

, & Altman,

D.G.

(2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Medicine, 6(7) https://doi.org/10.1371/journal.pmed.1000097

49.

Morkisch,

, & Dohle,

(2015). BeST - Berliner Spiegeltherapieprotokoll - Ein wissenschaftlich evaluiertes Manual zur Durchführung der Spiegeltherapie. Bad Honnef: Hippocampus Verlag.

50.

Morkisch,

, Jettkowski,

, Ferreira dos Santos,

, & Dohle,

(2017a). Modelling of therapeutic action during mirror therapy- A qualitative analysis as basis for technical solutions. Current Directions in Biomedical Engineering, 3, 45–48. https://doi.org/DOI 10.1515/cdbme-2017-0010

51.

Morkisch,

, Lommack,

, Kadow,

, Rietz,

, & Dohle,

(2017b). Evaluation eines standardisierten Therapieprotokolls zur Spiegeltherapie. Neurologie & Rehabilitation, 23(3), 216–226.

52.

Moustapha,

, & Rousseaux,

(2012). Immediate effects of mirror therapy on spatial neglect.. Annals of Physical and Rehabilitation Medicine, 55(S1), e197.

53.

Nagapattinam,

, Babu,

K.V.

, Kumar,

N.S.

, & Ayyappan,

V.R.

(2015). Effect of task specific mirror therapy with functional electrical stimulation on upper limb function for subacute hemiplegia. International Journal of Physiotherapy, 2(5) https://doi.org/10.15621/ijphy/2015/v2i5/78243

54.

Pandian,

, Arora,

, Kaur,

, Sharma,

, Vishwambaran,

, & Arima,

(2014). Mirror therapy in unilateral neglect after stroke (MUST trial): A randomized controlled trial. Neurology, 83(11), 1012–1017. https://doi.org/10.1212/WNL.0000000000000773

55.

Park,

, Chang,

, Kim,

K.-M

, & An,

D.-H.

(2015). The effects of mirror therapy with tasks on upper extremity function and self-care in stroke patients. Journal of Physical Therapy Science, 27(5), 1499–1501. https://doi.org/10.1589/jpts.27.1499

56.

Piravej,

, Champaiboon,

, Sontim,

, & Ruengyoo,

(2012). Effect of mirror therapy in recovering upper limb strength and function in chronic stroke patients.. Neurorehabilitation and Neural Repair, 26(6), Abst 126.

57.

Pollock,

, Farmer,

S.E

, Brady,

M.C

, Langhorne,

, Mead,

G.E

, Mehrholz,

, & van Wijck,

(2014). Interventions for improving upper limb function after stroke. The Cochrane Database of Systematic Reviews, (11) CD010820 https://https://doi.org/10.1002/14651858.CD010820.pub2

58.

Ramachandran,

V.S

, & Altschuler,

E.L.

(2009). The use of visual feedback, in particular mirror visual feedback, in restoring brain function. Brain, 132(7), 1693–1710. https://doi.org/10.1093/brain/awp135

59.

Review Manager (RevMan) (Version 5.3.). (2014). Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration.

60.

Rodrigues,

L.C

, Farias,

N.C

, Gomes,

R.P

, & Michaelsen,

S.M.

(2016). Feasibility and effectiveness of adding object-related bilateral symmetrical training to mirror therapy in chronic stroke: A randomized controlled pilot study. Physiotherapy Theory and Practice, 32(2), 83–91. https://doi.org/10.3109/09593985.2015.1091872

61.

Rothgangel,

, & Braun,

(2013). Mirror therapy. Practical protocol for stroke rehabilitation. https://doi.org/10.12855/ar.sb.mirrortherapy.e2013

62.

Samuelkamaleshkumar,

, Reethajanetsureka,

, Pauljebaraj,

, Benshamir,

, Padankatti,

S.M

, & David,

J.A.

(2014). Mirror therapy enhances motor performance in the paretic upper limb after stroke: A pilot randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 95(11), 2000–2005. https://doi.org/10.1016/j.apmr.2014.06.020

63.

Sanchez-Vives,

M.V

, & Slater,

(2005). From presence to consciousness through virtual reality. Nature Reviews. Neuroscience, 6(4), 332–339. https://doi.org/10.1038/nrn1651

64.

Sawner,

, & LaVigne,

(1992). Brunnstrom’s Movement Therapy in Hemiplegia: A Neurophysological Approach. (Philadelphia, PA: J.B. Lippincott).

65.

Schick,

, Schlake,

H.-P

, Kallusky,

, Hohlfeld,

, Steinmetz,

, Tripp,

, & Dohle,

(2017). Synergy effects of combined multichannel EMG-triggered electrical stimulation and mirror therapy in subacute stroke patients with severe or very severe arm/hand paresis. Restorative Neurology and Neuroscience, 35(3), 319–332. https://doi.org/10.3233/RNN-160710

66.

Seifert,

, Jettkowski,

, Pawelke,

, Morkisch,

, Schübl,

, & Dohle,

(2017). Analyse der Routineanwendung der Spiegeltherapie nach Schlaganfall. Neurol Rehabil, 23, S50 Berlin: Hippocampus Verlag.

67.

Selles,

R.W

, Michielsen,

M.E

, Bussmann,

J. B. J

, Stam,

H.J

, Hurkmans,

, Heijnen,

, & Ribbers,

G.M.

(2014). Effects of a mirror-induced visual illusion on a reaching task in stroke patients: Implications for mirror therapy training. Neurorehabilitation and Neural Repair 9 https://https://doi.org/10.1177/1545968314521005

68.

Seok,

, Kim,

, Jang,

, Lee,

, & Kim,

(2010). Effect of mirror therapy on recovery of upper limb function and strength in subacute hemiplegia after stroke. Journal of Korean Academy of Rehabilitation Medicine, 34, 508–512.

69.

Sütbeyaz,

, Yavuzer,

, Sezer,

, & Koseoglu,

B.F.

(2007). Mirror therapy enhances lower-extremity motor recovery and motor functioning after stroke: A randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 88(5), 555–559. https://doi.org/10.1016/j.apmr.2007.02.034

70.

Tezuka,

, Fujiwara,

, Kikuchi,

, Ogawa,

, Tokunaga,

, Ichikawa,

, & Katsuyama,

(2006). Effect of mirror therapy for patients with post-stroke paralysis of upper limb: Randomized cross-over study. Journal of Japanese Physical Therapy Association, 33(2), 62–68. https://doi.org/10.15063/rigaku.KJ00004317391

71.

Thieme,

, Bayn,

, Wurg,

, Zange,

, Pohl,

, & Behrens,

(2013). Mirror therapy for patients with severe arm paresis after stroke - a randomized controlled trial. Clinical Rehabilitation, 27(4), 314–324. https://doi.org/10.1177/0269215512455651

72.

Thieme,

, Morkisch,

, Mehrholz,

, Pohl,

, Behrens,

, Borgetto,

, & Dohle,

(2018). Mirror therapy for improving motor function after stroke. The Cochrane Database of Systematic Reviews, 7, CD008449 https://doi.org/10.1002/14651858.CD008449.pub3

73.

Tyson,

, Wilkinson,

, Thomas,

, Selles,

, McCabe,

, Tyrrell,

, & Vail,

(2015). Phase II pragmatic randomized controlled trial of patient-led therapies (mirror therapy and lower-limb exercises) during inpatient stroke rehabilitation. Neurorehabilitation and Neural Repair, 29(9), 818–826. https://doi.org/10.1177/1545968314565513

74.

Veerbeek,

J.M

, van Wegen,

, van Peppen,

, van der Wees,

P.J

, Hendriks,

, Rietberg,

, & Kwakkel,

(2014). What is the evidence for physical therapy poststroke? A systematic review and meta-analysis. PLoS ONE, 9(2) https://doi.org/10.1371/journal.pone.0087987

75.

Waltemate,

, Gall,

, Roth,

, Botsch,

, & Latoschik,

M.E.

(2018). The impact of avatar personalization and immersion on virtual body ownership, presence, and emotional response. IEEE Transactions on Visualization and Computer Graphics, 24(4), 1643–1652. https://doi.org/10.1109/TVCG.2018.2794629

76.

Wang,

, Fritzsch,

, Bernarding,

, Krause,

, Mauritz,

K.-H

, Brunetti,

, & Dohle,

(2013). Cerebral activation evoked by the mirror illusion of the hand in stroke patients compared to normal subjects. NeuroRehabilitation 593–603. https://doi.org/10.3233/NRE-130999

77.

Wang,

L.-J

, Chen,

L.-Z

, Ou,

, Guo,

, Hao,

, Chen,

, & Hu,

W.-L.

(2015). Effects of mirror visual feedback and electromyographic biofeedback on upper extremity function in hemiplegics after stroke. Chinese Journal of Rehabilitation Theory and Practice, 21(2), 202–206.

78.

Wu,

C.-Y

, Huang,

P.-C

, Chen,

Y.-T

, Lin,

K.-C

, & Yang,

H.-W.

(2013). Effects of mirror therapy on motor and sensory recovery in chronic stroke: A randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 94(6), 1023–1030. https://doi.org/10.1016/j.apmr.2013.02.007

79.

Yavuzer,

, Selles,

, Sezer,

, Sütbeyaz,

, Bussmann,

J.B

, Köseoğlu,

, . . . Stam,

H.J.

(2008). Mirror therapy improves hand function in subacute stroke: A randomized controlled trial. Archives of Physical Medicine and Rehabilitation, 89(3), 393–398. https://doi.org/10.1016/j.apmr.2007.08.162

80.

Yoon,

J.A

, Koo,

B.I

, Shin,

M.J

, Shin,

Y.B

, Ko,

H.-Y

, & Shin,

Y.-I.

(2014). Effect of constraint-induced movement therapy and mirror therapy for patients with subacute stroke. Annals of Rehabilitation Medicine, 38(4), 458 https://doi.org/10.5535/arm.2014.38.4.458

81.

Yun,

G.J

, Chun,

M.H

, Park,

J.Y

, & Kim,

B.R.

(2011). The synergic effects of mirror therapy and neuromuscular electrical stimulation for hand function in stroke patients. Annals of Rehabilitation Medicine, 35(3), 316 https://doi.org/10.5535/arm.2011.35.3.316